Generalization in Human Category Learning: A Connectionist Account of Differences in Gradient after Discriminative and Non discriminative Training

Two experiments are reported that investigate the difference in gradient of generalization observed between one-category (non-discriminative) and two-category (discriminative) training. Extrapolating from the results of a number of animal learning studies, it was predicted that the gradient should be steeper under discriminative training. The first experiment confirms this basic prediction for the stimuli used, which were novel, prototype-structured, and constructed from 12 symbols positioned on a grid. An explanation for the effect, based on the Rescorla-Wagner theory of Pavlovian conditioning (Rescorla & Wagner, 1972), is that under non-discriminative training 'incidental stimuli' have significant control over responding,whereas under discriminative training they do not. Incidental stimuli are those aspects of the stimulus, or the surrounding context, that are not differentially reinforced under discriminative training. This explanation leads to the prediction that a comparable effect of blocked versus intermixed discriminative training should also be found. This prediction is disconfirmed by the second experiment. An alternative model, still based on the RescorlaWagner theory but with the addition of a decision mechanism comprising a threshold unit and a competitive network system, is proposed, and its ability to predict both the choice probabilities and the pattern of response times found is evaluated via simulation.     

Circular vection as a function of the relative sizes, distances, and positions of two competing visual displays

In studies where it is reported that illusory self-rotation (circular vection) is induced more by peripheral displays than by central displays, eccentricity may have been confounded with perceived relative distance and area. Experiments are reported in which the direction and magnitude of vection induced by a central display in the presence of a surround display were measured. The displays varied in relative distance and area and were presented in isolation, with one moving and one stationary display, or with both moving in opposite directions. A more distant display had more influence over vection than a near display. A central display induced vection if seen in isolation or through a 'window' in a stationary surrounding display. Motion of a more distant central display weakened vection induced by a nearer surrounding display moving the other way. When the two displays had the same area their effects almost cancelled. A moving central display nearer than a textured stationary surround produced vection in the same direction as the moving stimulus. This phenomenon is termed 'contrast-motion vecton' because it is probably due to illusory motion of the surround induced by motion of the centre. Unequivocal statements about the dominance of an eccentric display over a central display cannot be made without considering the relative distances and sizes of the displays and the motion contrast between them.     

Accuracy of health-related quality of life assessment: what is the benefit of incorporating patients' preferences for domain functioning?

This study investigated the accuracy benefit of incorporating patients' preferences for domains of functioning into health-related quality of life (HRQOL) measurement. Using policy-capturing techniques, 102 medical outpatients judged the HRQOL of 16 scenarios describing varying levels of functioning in 3 domains. For each participant, regression analysis determined relative domain preferences and 2 decision models were built: one incorporating (preference-weighted) and one ignoring (equally weighted) domain preferences. To assess accuracy, the average proportion of judgment variance accounted for by each model was determined and both accounted for approximately 50%. However, for patients showing the greatest differences in importance across domains, the preference-weighted model was more accurate. Findings are discussed in the context of enhancing HRQOL assessment.     

Individual differences in perceptual sensitivity and response bias in anxiety: Evidence from emotional faces

We investigated the perception of emotional stimuli in anxious individuals and non-anxious cohorts. Signal detection theory analysis was applied to the discrimination of emotionally charged faces at several points along a continuum of emotional intensity. This design permitted the derivation of multiple measures of sensitivity and response bias for fearful and for happy faces. Anxious individuals lacked a conservative bias in judging fearful stimuli and a liberal bias in judging positive stimuli compared with non-anxious individuals. In addition, anxious participants had lower perceptual sensitivity (d′) than non-anxious participants for mildly threatening stimuli, as well as a trend towards lower perceptual sensitivity for moderately positive stimuli. These results suggest that the processing of threat information in anxiety is affected by sensitivity and bias differently at different levels of affective intensity.     

Perceptual learning in humans: Roles of preexposure schedule, feedback, and discrimination assay

In three experiments, humans received preexposure to two compound flavours (AX and BX: saline-lemon and sucrose-lemon) that were presented either in an intermixed (e.g., AX, BX, . . . BX, AX, . . .) or a blocked (e.g., AX, AX, . . . BX, BX. . .) fashion. Subsequently, AXwas paired with an unpleasant bitter taste, and the discriminability of AX and BX was assessed using the accuracy of same/different judgements and by the extent to which any learned dislike of AX generalized to BX. When participants received feedback about the accuracy of their same/different judgements during preexposure those given intermixed preexposure were more accurate in making these judgements during the test than those given blocked preexposure (Experiments 1 and 2A), however, there was no evidence of any learned dislike in these experiments. In Experiment 2B, in which participants did not receive feedback about the accuracy of their judgements, there was no effect of the preexposure regime on same/different judgements, but there was a learned dislike of AX, and this generalized less to BX in participants given intermixed than in those given blocked preexposure. The beneficial effects of intermixed preexposure are consistent with results from other species (chicks and rats), but the differences created by the presence or absence of feedback place constraints on the analysis of these effects.     

Behavioral Risk Assessment of the Guarded Suicidal Patient

Psychiatrists and other mental health professionals are trained to assess patients by direct observation and examination. Short inpatient length of stay, brief outpatient visits, emergency room evaluations, and other time‐limited clinical settings require rapid assessment of suicide risk. Recognition of behavioral suicide risk factors can assist in the early identification of the guarded suicidal patient, thus avoiding total reliance on the patient's reporting.